| Solution Feature |
- The whole system contains a Signal Collection of motor, fans, rubber belt and outlet parameters; Data Transfer from edge device to server; Data Preprocessing like wavelet analysis; AI model Predictions and Dashboards for the operator
- The raw data is measured by various sensors. Including current, voltage, shaft vibration, temperature of shaft/stator, velocity and stator vibration of the motor; velocity, vibration and temperature of the fan axle; temperature and vibration velocity of the rubber belt; air velocity, air pressure difference of the outlet
- After sensors are installed, raw data will be collected and preprocessed by traditional electrical engineering methods, such as wave packet analysis. The clean data will be stored for AI model training. A custom designed AI model will be trained on the daily raw data
- As there won’t be enough failures to train the AI model. System failures will be deliberately produced under typical scenarios, such as bearing wear, rubber belt ageing, fan wear, etc. Trained by those data, the AI model will achieve the ability of failure prediction
- The AI model will be deployed on the cloud or the edge device. After receiving cleaned data, a real-time failure prediction analysis will be shown on the dashboard. An alarm will be raised if any unusual issue happens
|
